Where AI Gets Its Information About Your Brand

When someone asks ChatGPT, Gemini, or Claude about a product category, the response these models provide doesn’t come from nowhere. In many cases, it originates from specific pages that the model found during a search, and often these pages are cited in the response itself.

This changes the question worth asking. Instead of just asking, "Did my brand appear in the response?", the more useful question is, "Where did this response get its information about my market?".

Types of Sources That Feed This Response

Comparison pages, industry directories, specialized publications, forums, and online communities, press articles. These are the types of sources that typically appear when a model cites something in response to a specific market inquiry.

Appearing in the Response vs. Being the Source

A brand can appear in a response without its own website being cited as the source. This happens when the model used a third-party comparison page, an industry directory, or a press article to form that mention, rather than the brand's own site.

This detail matters because it changes what is worth investigating. If the brand appears but never as its own source, the point of focus is not the mention itself, but who is telling the story in place of the brand.

How to Map These Sources

The approach is to repeat prompts that describe a buyer's need, without mentioning the brand name, in separate conversations on ChatGPT, Gemini, Claude, or Perplexity, and note the domains cited in each response. It’s worth repeating the same prompt multiple times: a study by SparkToro in partnership with Gumshoe.ai, with 2,961 prompt executions, found less than a 1 in 100 chance that two executions of the same prompt would yield exactly the same list of brands. The same type of variation applies to the cited sources.

After several executions, grouping the sources by frequency shows where the conversation about that market is happening, from the AI's perspective.

Two Ways to Claim Space in These Sources

With the most frequently cited sources mapped, there are two paths. The first is legitimate presence in them: updating a profile in a relevant directory, contributing to a specialized publication, participating in real discussions in communities where market buyers are already seeking recommendations. The second is creating your own content: if a competitor has a comparison page that appears as a source and the brand does not have an equivalent, that is a concrete gap to fill on its own site.

How MencionAI Tracks This

At MencionAI, source tracking is part of the regular monitoring of each prompt: the platform records which domains and pages the models cite or use as references when responding, prompt by prompt and model by model, over time. This eliminates the manual work of repeating and noting each execution, and maintains a history to compare if a new source has started appearing, or if a source that previously cited the brand has stopped appearing.

The Complete Process in One Guide

This text covers the central idea. The complete step-by-step process, with criteria for manually identifying sources and what to do with what is found, is in the guide How to Discover Which Sources AIs Use to Talk About Your Brand?.

Want to know which sources are citing your category right now?

Schedule a demo

Pedro, CMO of MencionAI

Frequently Asked Questions

Does every AI response show the sources used? No. Perplexity cites sources by default. ChatGPT and Claude show citations when using web search, which depends on the question.

Does my brand need to be the cited source to appear in a response? Not necessarily. The brand can appear in a response even when the cited source is a third-party page, such as a comparison or an industry directory.

Does MencionAI automatically show these sources? Yes. The monitoring of each prompt includes the recording of which domains and pages the models cite, without needing to repeat the test manually.